SAS output
- 글쓴이
- Simon
- 등록일
- 2004-12-14 04:37
- 조회
- 5,102회
- 추천
- 34건
- 댓글
- 0건
관련링크
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 27
The GLM Procedure
Class Level Information
Class Levels Values
year 6 1999 2000 2001 2002 2003 2004
media 4 YONHAP Chosun Dong-a Joongang
Number of observations 24
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 28
The GLM Procedure
Dependent Variable: number
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 8 1493.000000 186.625000 27.19 <.0001
Error 15 102.958333 6.863889
Corrected Total 23 1595.958333
R-Square Coeff Var Root MSE number Mean
0.935488 31.91759 2.619902 8.208333
Source DF Type I SS Mean Square F Value Pr > F
year 5 1437.208333 287.441667 41.88 <.0001
media 3 55.791667 18.597222 2.71 0.0822
Source DF Type III SS Mean Square F Value Pr > F
year 5 1437.208333 287.441667 41.88 <.0001
media 3 55.791667 18.597222 2.71 0.0822
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 29
The GLM Procedure
Level of ------------number-----------
year N Mean Std Dev
1999 4 5.2500000 3.20156212
2000 4 4.5000000 1.00000000
2001 4 4.5000000 1.29099445
2002 4 5.0000000 2.44948974
2003 4 4.5000000 0.57735027
2004 4 25.5000000 5.80229840
Level of ------------number-----------
media N Mean Std Dev
YONHAP 6 9.00000000 11.3313724
Chosun 6 8.50000000 8.2401456
Dong-a 6 9.66666667 8.6409876
Joongang 6 5.66666667 6.0882400
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 30
The GLM Procedure
t Tests (LSD) for number
NOTE: This test controls the Type I comparisonwise error rate, not the experimentwise
error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of t 2.13145
Least Significant Difference 3.9486
Means with the same letter are not significantly different.
t Grouping Mean N year
A 25.500 4 2004
B 5.250 4 1999
B
B 5.000 4 2002
B
B 4.500 4 2001
B
B 4.500 4 2003
B
B 4.500 4 2000
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 31
The GLM Procedure
Duncan's Multiple Range Test for number
NOTE: This test controls the Type I comparisonwise error rate, not the experimentwise
error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Number of Means 2 3 4 5 6
Critical Range 3.949 4.139 4.258 4.338 4.396
Means with the same letter are not significantly different.
Duncan Grouping Mean N year
A 25.500 4 2004
B 5.250 4 1999
B
B 5.000 4 2002
B
B 4.500 4 2001
B
B 4.500 4 2003
B
B 4.500 4 2000
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 32
The GLM Procedure
Tukey's Studentized Range (HSD) Test for number
NOTE: This test controls the Type I experimentwise error rate, but it generally has a
higher Type II error rate than REGWQ.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of Studentized Range 4.59474
Minimum Significant Difference 6.0189
Means with the same letter are not significantly different.
Tukey Grouping Mean N year
A 25.500 4 2004
B 5.250 4 1999
B
B 5.000 4 2002
B
B 4.500 4 2001
B
B 4.500 4 2003
B
B 4.500 4 2000
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 33
The GLM Procedure
Scheffe's Test for number
NOTE: This test controls the Type I experimentwise error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of F 2.90129
Minimum Significant Difference 7.0559
Means with the same letter are not significantly different.
Scheffe Grouping Mean N year
A 25.500 4 2004
B 5.250 4 1999
B
B 5.000 4 2002
B
B 4.500 4 2001
B
B 4.500 4 2003
B
B 4.500 4 2000
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 34
The GLM Procedure
t Tests (LSD) for number
NOTE: This test controls the Type I comparisonwise error rate, not the experimentwise
error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of t 2.13145
Least Significant Difference 3.224
Means with the same letter are not significantly different.
t Grouping Mean N media
A 9.667 6 Dong-a
A
A 9.000 6 YONHAP
A
B A 8.500 6 Chosun
B
B 5.667 6 Joongang
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 35
The GLM Procedure
Duncan's Multiple Range Test for number
NOTE: This test controls the Type I comparisonwise error rate, not the experimentwise
error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Number of Means 2 3 4
Critical Range 3.224 3.380 3.476
Means with the same letter are not significantly different.
Duncan Grouping Mean N media
A 9.667 6 Dong-a
A
B A 9.000 6 YONHAP
B A
B A 8.500 6 Chosun
B
B 5.667 6 Joongang
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 36
The GLM Procedure
Tukey's Studentized Range (HSD) Test for number
NOTE: This test controls the Type I experimentwise error rate, but it generally has a
higher Type II error rate than REGWQ.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of Studentized Range 4.07597
Minimum Significant Difference 4.3595
Means with the same letter are not significantly different.
Tukey Grouping Mean N media
A 9.667 6 Dong-a
A
A 9.000 6 YONHAP
A
A 8.500 6 Chosun
A
A 5.667 6 Joongang
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 37
The GLM Procedure
Scheffe's Test for number
NOTE: This test controls the Type I experimentwise error rate.
Alpha 0.05
Error Degrees of Freedom 15
Error Mean Square 6.863889
Critical Value of F 3.28738
Minimum Significant Difference 4.7502
Means with the same letter are not significantly different.
Scheffe Grouping Mean N media
A 9.667 6 Dong-a
A
A 9.000 6 YONHAP
A
A 8.500 6 Chosun
A
A 5.667 6 Joongang
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 38
The GLM Procedure
Level of ------------number-----------
year N Mean Std Dev
1999 4 5.2500000 3.20156212
2000 4 4.5000000 1.00000000
2001 4 4.5000000 1.29099445
2002 4 5.0000000 2.44948974
2003 4 4.5000000 0.57735027
2004 4 25.5000000 5.80229840
Level of ------------number-----------
media N Mean Std Dev
YONHAP 6 9.00000000 11.3313724
Chosun 6 8.50000000 8.2401456
Dong-a 6 9.66666667 8.6409876
Joongang 6 5.66666667 6.0882400
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 39
The GLM Procedure
Level of ------------number-----------
year N Mean Std Dev
1999 4 5.2500000 3.20156212
2000 4 4.5000000 1.00000000
2001 4 4.5000000 1.29099445
2002 4 5.0000000 2.44948974
2003 4 4.5000000 0.57735027
2004 4 25.5000000 5.80229840
Level of ------------number-----------
media N Mean Std Dev
YONHAP 6 9.00000000 11.3313724
Chosun 6 8.50000000 8.2401456
Dong-a 6 9.66666667 8.6409876
Joongang 6 5.66666667 6.0882400
------------------------------------------------------------------------------------------
ANOVA on Media Plays by MOCIE for the Past 6 Years - Multiple Pairwise Comparison 40
The GLM Procedure
Dependent Variable: number
Contrast DF Contrast SS Mean Square F Value Pr > F
Linear contrast 1 591.6035714 591.6035714 86.19 <.0001
Quadratic contrast 1 542.6458333 542.6458333 79.06 <.0001
Cubic contrast 1 218.9013889 218.9013889 31.89 <.0001
Quartic contrast 1 73.9375000 73.9375000 10.77 0.0050
Quintic contrast 1 10.1200397 10.1200397 1.47 0.2434
------------------------------------------------------------------------------------------
첫째, 4대 신문의 기술유출관련 보도회수는 2004년 유독 증가했다.
둘째, 다소 보수적 견지를 고수(Tukey & Scheffe's)하여 판단컨데 네 개 신문사의 보도회수는 통계학적으로 큰 차이를 보이지 않는다. (즉, 해당년도 마다 비슷한 회수의 관련기사를 신문사마다 내어왔음)
세째, 기술유출 및 인력유출에 관한 구분이 필요하며, 기사의 양적 증가와 별도로 독자 및 정책결정자에 큰 영향을 줄 수 있는 기술유출 및 산업스파이 행위관련 보도내용의 수준, 즉 질적 상태에 관한 검증이 필요하다. 단순한 관련기사의 양적 증가보다, 1개의 사설, 소수 개의 기사일 지언정 마녀사냥식 여론몰이를 지향하는 강력한 의지가 포함된 파급효과 큰 기사의 경우, 다수 개의 양적 보도에 비해 훨씬 지대한 그릇된 여론 형성력을 지니고 있기 때문이다.
끝으로, 우선 4대 신문사의 지난 7년간 (혹은 지난 6년간) 기술유출 관련기사들 하나 하나 검토한 후 그 내용에 문제는 없었는지 추가 검토 작업을 수행할 필요성이 있으며 향후 단계적으로 시도키로 한다.